Labeling Graphs

Recently, WikiLeaks released a draft text of the intellectual property chapter of the Trans-Pacific Partnership agreement. In what is otherwise an excellent (and somewhat disturbing) analysis of the data done by Gabriel Michael, a Ph.D. candidate at George Washington University, there is one problem: I couldn’t read the labeling on the graphs very easily. In this case, since the data is heavily tied to connections between countries, it’s essential to understanding the relationships presented in the graphs.

To a degree, this was exacerbated by the layout of the Washington Post’s article template ( which is where I first viewed these. But even viewed on the author’s original site – somewhat larger – they can still be difficult to read (


Allow me to make it clear that I’m not picking on anybody here or denigrating some very fine scholarship. I’m just adding a couple of modest improvements in the hope that graphics such as these will be able to reach an even wider audience. They carry a message that really needs to be received by as many people as possible.

First, the obvious. I increased the font size. However, it’s (usually) more legible to also change to a sans-serif typeface…and, while we’re at it, let’s bump up the weight. So now we have:


It’s worth noting that I was able to save the image at a smaller size too. It’s 440 pixels wide (including a bit of white space) and could now conceivably be viewed on a phone.

There is another improvement that may be made: Getting rid of the thin black lines around the circles. They don’t fit with the rest of the imagery in the graphs. Furthermore, a hollow shape requires a bit more visual decoding by the viewer’s eyes. Of course, removing those lines also means reversing the text color and picking some solid color for the circles. I chose a grayish-blue, not too saturated, to provide a bit of contrast to the yellow-orange-red lines:


Again, this post is not meant to be anything more than constructive. We all have our strengths and weaknesses.